in_pcb.c revision 1.25
1 /* $NetBSD: in_pcb.c,v 1.25 1996/02/13 23:41:53 christos Exp $ */ 2 3 /* 4 * Copyright (c) 1982, 1986, 1991, 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. All advertising materials mentioning features or use of this software 16 * must display the following acknowledgement: 17 * This product includes software developed by the University of 18 * California, Berkeley and its contributors. 19 * 4. Neither the name of the University nor the names of its contributors 20 * may be used to endorse or promote products derived from this software 21 * without specific prior written permission. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33 * SUCH DAMAGE. 34 * 35 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94 36 */ 37 38 #include <sys/param.h> 39 #include <sys/systm.h> 40 #include <sys/malloc.h> 41 #include <sys/mbuf.h> 42 #include <sys/protosw.h> 43 #include <sys/socket.h> 44 #include <sys/socketvar.h> 45 #include <sys/ioctl.h> 46 #include <sys/errno.h> 47 #include <sys/time.h> 48 #include <sys/proc.h> 49 50 #include <net/if.h> 51 #include <net/route.h> 52 53 #include <netinet/in.h> 54 #include <netinet/in_systm.h> 55 #include <netinet/ip.h> 56 #include <netinet/in_pcb.h> 57 #include <netinet/in_var.h> 58 #include <netinet/ip_var.h> 59 60 struct in_addr zeroin_addr; 61 62 #define INPCBHASH(table, faddr, fport, laddr, lport) \ 63 &(table)->inpt_hashtbl[(ntohl((faddr)->s_addr) + ntohs((fport)) + ntohs((lport))) & (table->inpt_hash)] 64 65 void 66 in_pcbinit(table, hashsize) 67 struct inpcbtable *table; 68 int hashsize; 69 { 70 71 CIRCLEQ_INIT(&table->inpt_queue); 72 table->inpt_hashtbl = hashinit(hashsize, M_PCB, &table->inpt_hash); 73 table->inpt_lastport = 0; 74 } 75 76 int 77 in_pcballoc(so, v) 78 struct socket *so; 79 void *v; 80 { 81 struct inpcbtable *table = v; 82 register struct inpcb *inp; 83 int s; 84 85 MALLOC(inp, struct inpcb *, sizeof(*inp), M_PCB, M_WAITOK); 86 if (inp == NULL) 87 return (ENOBUFS); 88 bzero((caddr_t)inp, sizeof(*inp)); 89 inp->inp_table = table; 90 inp->inp_socket = so; 91 s = splnet(); 92 CIRCLEQ_INSERT_HEAD(&table->inpt_queue, inp, inp_queue); 93 LIST_INSERT_HEAD(INPCBHASH(table, &inp->inp_faddr, inp->inp_fport, 94 &inp->inp_laddr, inp->inp_lport), inp, inp_hash); 95 splx(s); 96 so->so_pcb = inp; 97 return (0); 98 } 99 100 int 101 in_pcbbind(v, nam) 102 register void *v; 103 struct mbuf *nam; 104 { 105 register struct inpcb *inp = v; 106 register struct socket *so = inp->inp_socket; 107 register struct inpcbtable *table = inp->inp_table; 108 register struct sockaddr_in *sin; 109 struct proc *p = curproc; /* XXX */ 110 u_int16_t lport = 0; 111 int wild = 0, reuseport = (so->so_options & SO_REUSEPORT); 112 int error; 113 114 if (in_ifaddr.tqh_first == 0) 115 return (EADDRNOTAVAIL); 116 if (inp->inp_lport || inp->inp_laddr.s_addr != INADDR_ANY) 117 return (EINVAL); 118 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0 && 119 ((so->so_proto->pr_flags & PR_CONNREQUIRED) == 0 || 120 (so->so_options & SO_ACCEPTCONN) == 0)) 121 wild = INPLOOKUP_WILDCARD; 122 if (nam) { 123 sin = mtod(nam, struct sockaddr_in *); 124 if (nam->m_len != sizeof (*sin)) 125 return (EINVAL); 126 #ifdef notdef 127 /* 128 * We should check the family, but old programs 129 * incorrectly fail to initialize it. 130 */ 131 if (sin->sin_family != AF_INET) 132 return (EAFNOSUPPORT); 133 #endif 134 lport = sin->sin_port; 135 if (IN_MULTICAST(sin->sin_addr.s_addr)) { 136 /* 137 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; 138 * allow complete duplication of binding if 139 * SO_REUSEPORT is set, or if SO_REUSEADDR is set 140 * and a multicast address is bound on both 141 * new and duplicated sockets. 142 */ 143 if (so->so_options & SO_REUSEADDR) 144 reuseport = SO_REUSEADDR|SO_REUSEPORT; 145 } else if (sin->sin_addr.s_addr != INADDR_ANY) { 146 sin->sin_port = 0; /* yech... */ 147 if (ifa_ifwithaddr(sintosa(sin)) == 0) 148 return (EADDRNOTAVAIL); 149 } 150 if (lport) { 151 struct inpcb *t; 152 153 /* GROSS */ 154 if (ntohs(lport) < IPPORT_RESERVED && 155 (error = suser(p->p_ucred, &p->p_acflag))) 156 return (EACCES); 157 t = in_pcblookup(table, zeroin_addr, 0, 158 sin->sin_addr, lport, wild); 159 if (t && (reuseport & t->inp_socket->so_options) == 0) 160 return (EADDRINUSE); 161 } 162 inp->inp_laddr = sin->sin_addr; 163 } 164 if (lport == 0) 165 do { 166 if (table->inpt_lastport++ < IPPORT_RESERVED || 167 table->inpt_lastport > IPPORT_USERRESERVED) 168 table->inpt_lastport = IPPORT_RESERVED; 169 lport = htons(table->inpt_lastport); 170 } while (in_pcblookup(table, 171 zeroin_addr, 0, inp->inp_laddr, lport, wild)); 172 inp->inp_lport = lport; 173 in_pcbrehash(inp); 174 return (0); 175 } 176 177 /* 178 * Connect from a socket to a specified address. 179 * Both address and port must be specified in argument sin. 180 * If don't have a local address for this socket yet, 181 * then pick one. 182 */ 183 int 184 in_pcbconnect(v, nam) 185 register void *v; 186 struct mbuf *nam; 187 { 188 register struct inpcb *inp = v; 189 struct in_ifaddr *ia; 190 struct sockaddr_in *ifaddr = NULL; 191 register struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *); 192 193 if (nam->m_len != sizeof (*sin)) 194 return (EINVAL); 195 if (sin->sin_family != AF_INET) 196 return (EAFNOSUPPORT); 197 if (sin->sin_port == 0) 198 return (EADDRNOTAVAIL); 199 if (in_ifaddr.tqh_first != 0) { 200 /* 201 * If the destination address is INADDR_ANY, 202 * use the primary local address. 203 * If the supplied address is INADDR_BROADCAST, 204 * and the primary interface supports broadcast, 205 * choose the broadcast address for that interface. 206 */ 207 if (sin->sin_addr.s_addr == INADDR_ANY) 208 sin->sin_addr = in_ifaddr.tqh_first->ia_addr.sin_addr; 209 else if (sin->sin_addr.s_addr == INADDR_BROADCAST && 210 (in_ifaddr.tqh_first->ia_ifp->if_flags & IFF_BROADCAST)) 211 sin->sin_addr = in_ifaddr.tqh_first->ia_broadaddr.sin_addr; 212 } 213 if (inp->inp_laddr.s_addr == INADDR_ANY) { 214 register struct route *ro; 215 216 ia = (struct in_ifaddr *)0; 217 /* 218 * If route is known or can be allocated now, 219 * our src addr is taken from the i/f, else punt. 220 */ 221 ro = &inp->inp_route; 222 if (ro->ro_rt && 223 (satosin(&ro->ro_dst)->sin_addr.s_addr != 224 sin->sin_addr.s_addr || 225 inp->inp_socket->so_options & SO_DONTROUTE)) { 226 RTFREE(ro->ro_rt); 227 ro->ro_rt = (struct rtentry *)0; 228 } 229 if ((inp->inp_socket->so_options & SO_DONTROUTE) == 0 && /*XXX*/ 230 (ro->ro_rt == (struct rtentry *)0 || 231 ro->ro_rt->rt_ifp == (struct ifnet *)0)) { 232 /* No route yet, so try to acquire one */ 233 ro->ro_dst.sa_family = AF_INET; 234 ro->ro_dst.sa_len = sizeof(struct sockaddr_in); 235 satosin(&ro->ro_dst)->sin_addr = sin->sin_addr; 236 rtalloc(ro); 237 } 238 /* 239 * If we found a route, use the address 240 * corresponding to the outgoing interface 241 * unless it is the loopback (in case a route 242 * to our address on another net goes to loopback). 243 */ 244 if (ro->ro_rt && !(ro->ro_rt->rt_ifp->if_flags & IFF_LOOPBACK)) 245 ia = ifatoia(ro->ro_rt->rt_ifa); 246 if (ia == 0) { 247 u_int16_t fport = sin->sin_port; 248 249 sin->sin_port = 0; 250 ia = ifatoia(ifa_ifwithdstaddr(sintosa(sin))); 251 if (ia == 0) 252 ia = ifatoia(ifa_ifwithnet(sintosa(sin))); 253 sin->sin_port = fport; 254 if (ia == 0) 255 ia = in_ifaddr.tqh_first; 256 if (ia == 0) 257 return (EADDRNOTAVAIL); 258 } 259 /* 260 * If the destination address is multicast and an outgoing 261 * interface has been set as a multicast option, use the 262 * address of that interface as our source address. 263 */ 264 if (IN_MULTICAST(sin->sin_addr.s_addr) && 265 inp->inp_moptions != NULL) { 266 struct ip_moptions *imo; 267 struct ifnet *ifp; 268 269 imo = inp->inp_moptions; 270 if (imo->imo_multicast_ifp != NULL) { 271 ifp = imo->imo_multicast_ifp; 272 for (ia = in_ifaddr.tqh_first; ia != 0; 273 ia = ia->ia_list.tqe_next) 274 if (ia->ia_ifp == ifp) 275 break; 276 if (ia == 0) 277 return (EADDRNOTAVAIL); 278 } 279 } 280 ifaddr = satosin(&ia->ia_addr); 281 } 282 if (in_pcbhashlookup(inp->inp_table, sin->sin_addr, sin->sin_port, 283 inp->inp_laddr.s_addr ? inp->inp_laddr : ifaddr->sin_addr, 284 inp->inp_lport) != 0) 285 return (EADDRINUSE); 286 if (inp->inp_laddr.s_addr == INADDR_ANY) { 287 if (inp->inp_lport == 0) 288 (void)in_pcbbind(inp, (struct mbuf *)0); 289 inp->inp_laddr = ifaddr->sin_addr; 290 } 291 inp->inp_faddr = sin->sin_addr; 292 inp->inp_fport = sin->sin_port; 293 in_pcbrehash(inp); 294 return (0); 295 } 296 297 void 298 in_pcbdisconnect(v) 299 void *v; 300 { 301 struct inpcb *inp = v; 302 303 inp->inp_faddr.s_addr = INADDR_ANY; 304 inp->inp_fport = 0; 305 in_pcbrehash(inp); 306 if (inp->inp_socket->so_state & SS_NOFDREF) 307 in_pcbdetach(inp); 308 } 309 310 void 311 in_pcbdetach(v) 312 void *v; 313 { 314 struct inpcb *inp = v; 315 struct socket *so = inp->inp_socket; 316 int s; 317 318 so->so_pcb = 0; 319 sofree(so); 320 if (inp->inp_options) 321 (void)m_free(inp->inp_options); 322 if (inp->inp_route.ro_rt) 323 rtfree(inp->inp_route.ro_rt); 324 ip_freemoptions(inp->inp_moptions); 325 s = splnet(); 326 LIST_REMOVE(inp, inp_hash); 327 CIRCLEQ_REMOVE(&inp->inp_table->inpt_queue, inp, inp_queue); 328 splx(s); 329 FREE(inp, M_PCB); 330 } 331 332 void 333 in_setsockaddr(inp, nam) 334 register struct inpcb *inp; 335 struct mbuf *nam; 336 { 337 register struct sockaddr_in *sin; 338 339 nam->m_len = sizeof (*sin); 340 sin = mtod(nam, struct sockaddr_in *); 341 bzero((caddr_t)sin, sizeof (*sin)); 342 sin->sin_family = AF_INET; 343 sin->sin_len = sizeof(*sin); 344 sin->sin_port = inp->inp_lport; 345 sin->sin_addr = inp->inp_laddr; 346 } 347 348 void 349 in_setpeeraddr(inp, nam) 350 struct inpcb *inp; 351 struct mbuf *nam; 352 { 353 register struct sockaddr_in *sin; 354 355 nam->m_len = sizeof (*sin); 356 sin = mtod(nam, struct sockaddr_in *); 357 bzero((caddr_t)sin, sizeof (*sin)); 358 sin->sin_family = AF_INET; 359 sin->sin_len = sizeof(*sin); 360 sin->sin_port = inp->inp_fport; 361 sin->sin_addr = inp->inp_faddr; 362 } 363 364 /* 365 * Pass some notification to all connections of a protocol 366 * associated with address dst. The local address and/or port numbers 367 * may be specified to limit the search. The "usual action" will be 368 * taken, depending on the ctlinput cmd. The caller must filter any 369 * cmds that are uninteresting (e.g., no error in the map). 370 * Call the protocol specific routine (if any) to report 371 * any errors for each matching socket. 372 * 373 * Must be called at splsoftnet. 374 */ 375 void 376 in_pcbnotify(table, dst, fport_arg, laddr, lport_arg, errno, notify) 377 struct inpcbtable *table; 378 struct sockaddr *dst; 379 u_int fport_arg, lport_arg; 380 struct in_addr laddr; 381 int errno; 382 void (*notify) __P((struct inpcb *, int)); 383 { 384 register struct inpcb *inp, *oinp; 385 struct in_addr faddr; 386 u_int16_t fport = fport_arg, lport = lport_arg; 387 388 if (dst->sa_family != AF_INET) 389 return; 390 faddr = satosin(dst)->sin_addr; 391 if (faddr.s_addr == INADDR_ANY) 392 return; 393 394 for (inp = table->inpt_queue.cqh_first; 395 inp != (struct inpcb *)&table->inpt_queue;) { 396 if (inp->inp_faddr.s_addr != faddr.s_addr || 397 inp->inp_socket == 0 || 398 inp->inp_fport != fport || 399 inp->inp_lport != lport || 400 inp->inp_laddr.s_addr != laddr.s_addr) { 401 inp = inp->inp_queue.cqe_next; 402 continue; 403 } 404 oinp = inp; 405 inp = inp->inp_queue.cqe_next; 406 if (notify) 407 (*notify)(oinp, errno); 408 } 409 } 410 411 void 412 in_pcbnotifyall(table, dst, errno, notify) 413 struct inpcbtable *table; 414 struct sockaddr *dst; 415 int errno; 416 void (*notify) __P((struct inpcb *, int)); 417 { 418 register struct inpcb *inp, *oinp; 419 struct in_addr faddr; 420 421 if (dst->sa_family != AF_INET) 422 return; 423 faddr = satosin(dst)->sin_addr; 424 if (faddr.s_addr == INADDR_ANY) 425 return; 426 427 for (inp = table->inpt_queue.cqh_first; 428 inp != (struct inpcb *)&table->inpt_queue;) { 429 if (inp->inp_faddr.s_addr != faddr.s_addr || 430 inp->inp_socket == 0) { 431 inp = inp->inp_queue.cqe_next; 432 continue; 433 } 434 oinp = inp; 435 inp = inp->inp_queue.cqe_next; 436 if (notify) 437 (*notify)(oinp, errno); 438 } 439 } 440 441 /* 442 * Check for alternatives when higher level complains 443 * about service problems. For now, invalidate cached 444 * routing information. If the route was created dynamically 445 * (by a redirect), time to try a default gateway again. 446 */ 447 void 448 in_losing(inp) 449 struct inpcb *inp; 450 { 451 register struct rtentry *rt; 452 struct rt_addrinfo info; 453 454 if ((rt = inp->inp_route.ro_rt)) { 455 inp->inp_route.ro_rt = 0; 456 bzero((caddr_t)&info, sizeof(info)); 457 info.rti_info[RTAX_DST] = &inp->inp_route.ro_dst; 458 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 459 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 460 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0); 461 if (rt->rt_flags & RTF_DYNAMIC) 462 (void) rtrequest(RTM_DELETE, rt_key(rt), 463 rt->rt_gateway, rt_mask(rt), rt->rt_flags, 464 (struct rtentry **)0); 465 else 466 /* 467 * A new route can be allocated 468 * the next time output is attempted. 469 */ 470 rtfree(rt); 471 } 472 } 473 474 /* 475 * After a routing change, flush old routing 476 * and allocate a (hopefully) better one. 477 */ 478 void 479 in_rtchange(inp, errno) 480 register struct inpcb *inp; 481 int errno; 482 { 483 if (inp->inp_route.ro_rt) { 484 rtfree(inp->inp_route.ro_rt); 485 inp->inp_route.ro_rt = 0; 486 /* 487 * A new route can be allocated the next time 488 * output is attempted. 489 */ 490 } 491 } 492 493 struct inpcb * 494 in_pcblookup(table, faddr, fport_arg, laddr, lport_arg, flags) 495 struct inpcbtable *table; 496 struct in_addr faddr, laddr; 497 u_int fport_arg, lport_arg; 498 int flags; 499 { 500 register struct inpcb *inp, *match = 0; 501 int matchwild = 3, wildcard; 502 u_int16_t fport = fport_arg, lport = lport_arg; 503 504 for (inp = table->inpt_queue.cqh_first; 505 inp != (struct inpcb *)&table->inpt_queue; 506 inp = inp->inp_queue.cqe_next) { 507 if (inp->inp_lport != lport) 508 continue; 509 wildcard = 0; 510 if (inp->inp_faddr.s_addr != INADDR_ANY) { 511 if (faddr.s_addr == INADDR_ANY) 512 wildcard++; 513 else if (inp->inp_faddr.s_addr != faddr.s_addr || 514 inp->inp_fport != fport) 515 continue; 516 } else { 517 if (faddr.s_addr != INADDR_ANY) 518 wildcard++; 519 } 520 if (inp->inp_laddr.s_addr != INADDR_ANY) { 521 if (laddr.s_addr == INADDR_ANY) 522 wildcard++; 523 else if (inp->inp_laddr.s_addr != laddr.s_addr) 524 continue; 525 } else { 526 if (laddr.s_addr != INADDR_ANY) 527 wildcard++; 528 } 529 if (wildcard && (flags & INPLOOKUP_WILDCARD) == 0) 530 continue; 531 if (wildcard < matchwild) { 532 match = inp; 533 matchwild = wildcard; 534 if (matchwild == 0) 535 break; 536 } 537 } 538 return (match); 539 } 540 541 void 542 in_pcbrehash(inp) 543 struct inpcb *inp; 544 { 545 struct inpcbtable *table = inp->inp_table; 546 int s; 547 548 s = splnet(); 549 LIST_REMOVE(inp, inp_hash); 550 LIST_INSERT_HEAD(INPCBHASH(table, &inp->inp_faddr, inp->inp_fport, 551 &inp->inp_laddr, inp->inp_lport), inp, inp_hash); 552 splx(s); 553 } 554 555 #ifdef DIAGNOSTIC 556 int in_pcbnotifymiss = 0; 557 #endif 558 559 struct inpcb * 560 in_pcbhashlookup(table, faddr, fport_arg, laddr, lport_arg) 561 struct inpcbtable *table; 562 struct in_addr faddr, laddr; 563 u_int fport_arg, lport_arg; 564 { 565 struct inpcbhead *head; 566 register struct inpcb *inp; 567 u_int16_t fport = fport_arg, lport = lport_arg; 568 569 head = INPCBHASH(table, &faddr, fport, &laddr, lport); 570 for (inp = head->lh_first; inp != NULL; inp = inp->inp_hash.le_next) { 571 if (inp->inp_faddr.s_addr == faddr.s_addr && 572 inp->inp_fport == fport && 573 inp->inp_lport == lport && 574 inp->inp_laddr.s_addr == laddr.s_addr) { 575 /* 576 * Move this PCB to the head of hash chain so that 577 * repeated accesses are quicker. This is analogous to 578 * the historic single-entry PCB cache. 579 */ 580 if (inp != head->lh_first) { 581 LIST_REMOVE(inp, inp_hash); 582 LIST_INSERT_HEAD(head, inp, inp_hash); 583 } 584 break; 585 } 586 } 587 #ifdef DIAGNOSTIC 588 if (inp == NULL && in_pcbnotifymiss) { 589 printf("in_pcbhashlookup: faddr=%08x fport=%d laddr=%08x lport=%d\n", 590 ntohl(faddr.s_addr), ntohs(fport), 591 ntohl(laddr.s_addr), ntohs(lport)); 592 } 593 #endif 594 return (inp); 595 } 596
Indexes created Tue Sep 23 14:10:03 GMT 2025